Figur« S. Trapped noneondensable gas lowers heat transfer baffles can let coolant bypass of por" | tions of the tube bundle, reducing the ! coolant flow through the tubes and the : effective temperature difference for condensing. Common causes of exces-" ; sive clearance include corrosion, ero, j sion. and bowed tubesheets. The last i can be caused by excessive pressure oc. a thermally-induced tube contraction which can occur in a condenser having an inadequate shell expansion joint or ; one that is subjected to temperature-I differences between the shell and tubes i that exceed the design limits. ; Undersizing — A condenser that cannot handle the overhead vapor wilf operate at a higher-than-desired presj sure drop or with excessive venting. Sometimes, an inadequate heat-tran%_ fer area in a shellside liquid-cooled con denser can be compensated for bj changing the condenser's heads and pass-partitions to reduce the number of-tube passes, boosting the coolant flow and the temperature driving force. Shellside modifications to reduce the pressure drop are more difficult. be" cause baffle cuts and baffle and tubi spacing are difficult to alter.
Air leaking in through fittings and ves sel walls (due to the porosity of the metal) and noneondensable gases ei»-
tering in the feed stream can overload a vacuum system, diminishing condenser performance. Of course, a drop-off in condenser performance will also overload tile vacuum system with condensable vapors. Insufficient, sealant flow to a liquid-ring vacuum putnp can cause the pump to overheat and flash the seal fluid, curtailing the pump's capacity, as can worn seals and bearings. Diffuser throat erosion, intercondenser failure, inadequate cooling water, and low motive-steam pressure can undermine the performance of a vacuum steam jet.
An interruption in the operation of a jet or vacuum pump can result in the complete or partial loss of vacuum. A check valve or seal leg will prevent the loss of column vacuum due to vapor flowing back through the vent system. However, the column boilup will continue to push the normal inert load into the condenser, where it can quickly blanket the tubes and fill the exchanger. The column pressure will begin to
Downward vapor surge can collapse column trays rise to the limit capped by liquid boiling temperatures and the temperature of the re* boiler heaQQ£âiùi When returneoto service, a vacuum pump that may have taken hours to evacuate the inert gases from the column at startup will almost instantly evacuate the few cubic feet of inert gas blanketing the condenser surface area, restoring process condensing, almost immediately regaining the vacuum. This sudden drop in column pressure, compounded by the column having served as a heat sink at the higher, upset pressure, will produce an excessive boilup surge that can damage column internal structures. This surge can be minimized by manually opening the vacuum-control recycle stream before returning the vacuum pump or jet to service and slowly ramping to the desired pressure.
The case of the inconsistent column heat balance and control strategy
Problem: A packed column was designed to remove low-boiling components from much higher-boiling products for recycling to an upstream operation. Some of the products contained a varying mixture of high-boiling components (which caused the base temperature to fluctuate), the reboiler steam was flow controlled instead of base temperature controlled. Some high- or intermediate-temperature boiling components were allowed to pass overhead. Occasionally, the operator trimmed the steam flow to control the amount of high-boiling components taken overhead. At design feedrates, the column operated as expected, but it flooded when the feed was shut off.
Troubleshooting: Because the plant was new, the troubleshooter checked the design basis for the column. The design heat and material balances indicated that the feed was cold, requiring about 40% of the reboiler heat duty justtoheatthehigh-boilingcomponents in the feed to the underflow temperature. The column was sized for a vapor boilup corresponding to the remaining 60% of the reboiler dutv.
that flooding occurred when the feed was reduced or shut off only after operation near to the reboiler design steam flowrate. Because the reboiler steam was flow controlled, shutting off the feed removed the heat sink for 40% of the reboiler duty. The entire reboiler duty would then translate into boilup equivalent to 166% of the column design capacity. Thus, shutting off the feed eliminated a significant heat sink, causing excessive boilup into the column.
Corrective action: An upper limit to the column pressure drop was established, based on operation short of flooding. The distributed control system was modified to manipulate reboiler steam flow to maintain a column pressure drop below the limit. This would provide indirect steam flow control during normal operation (i.e., with feed), but would reduce the reboiler steam flow to curtail the higher boilup when the feed is reduced or shut off.
Outcome: Reducing the feed no longer started column flooding.
A sudden loss of vacuum from the top of a column may let vapor flow back through the column. Because of the presence of liquid on column trays, the vapor downflow will force liquid back through the trays at a high differential pressure. The force of this downflow can collapse trays (such as the bubble-cap tray in the photograph above! — and especially vaive trays, which by design close against downward flow.
Undersized or cavitating pumps will limit column capacity. Leaking seal fluid can introduce contaminants into the column. Pump discharge pressures should be checked periodically to ensure that impellers have not eroded or suction screens have not become plugged.
Pipes can be plugged by debris from a column (such as packing or loose valves) and process deposits. Vent lines may have low segments that can become filled with condensed vapors and cause backpressure problems. A field check with the piping and instrument drawings can help identify problems, such as inadequate liquid heads and seals, improper routings and open bypass valves.
Pressure changes in upstream or downstream columns, or other equipment. may limit pump capacity. The composition of the feed from a reactor to a column may vary, upsetting column performance. Unexpected or inconsistent byproduct formation can also undercut the column separation.
Variations in the temperatures of fluids from preheaters and interchang-ers may upset a column's heat balance. Overheating a liquid feed may flash it. and render it incompatible with the column's piping or internal arrangements. Reflux and feed decanters can become upset, and let impurities enter the column. Surges of low-boiling components (common if a column upstream becomes upset) can overwhelm the condenser and vacuum svstem.
Many sensing elements must be installed through nozzles located at specific elevations and orientations with respect to a distillation column's internal structures. For example, pressure transmitters are normally located in a vapor space. Similarly, temperature sensors for composition control should be placed by a tray that will provide sensitivity to composition changes. Sensors also should be designed and constructed to avoid corrosion, fouling and plugging.
Transmitters improperly calibrated will produce erroneous readings, which will create problems or confuse troubleshooting efforts. The troubleshooter should always consider the possibility that readings may be false and verify instrument measurements, especially those that deviate from normal.
An incorrectly calibrated level instrument is often the initiator of column flooding. For example, if an instrument for gauging the base level of a column has been calibrated for a liquid having a density of 64 lb/ft3 but is sensing a liquid having a density of 53 lb/ft"*, an indicated level of 4 ft would actually be over 4.8 ft. Thus, the instrument could raise the liquid level higher than the reboiler return nozzle and even into the column, causing entrainment and subsequently flooding.
A similar problem will occur if the liquid level is allowed to rise above the vapor tap. Because there will be a constant head of liquid between the two taps, any addition to the base level will not be detected. If the density of the calibration liquid is greater than that of the actual liquid (as in the foregoing example), the level transmitter will range out at less than 100%, conveying a misleading indication. This problem can be detected by noting a suspiciously constant level indication.
Transmitter legs that are open to process fluids often have to be purged or traced to keep liquid or condensate from accumulating in them and giving false readings. Some reasons why a control valve may not respond satisfactorily, besides being too large or too small, include having a sticking valve stem, an inadequate air supply, a malfunctioning valve positioner and a leaking diaphragm. ■
Additional reg^iog far this ioiur-part ti^yj^kehuoting report
1. Drew. J. \V.. Distillation CotatahiStarmp.
2. Frank. 0.. Shortcut* for Destination Design. Chem. Eng.. Mar. 14. 1977. pp. 1U-128.
Estimating Packed Coi Jcember 19JSTpp.
3. Hausch. D. C.. How Flooding Can Affect Tower Operation. Chem. Eng. Prog... October 1964. pp. 55-59.
4. Kister. H. Z.. How to PreDare and Test Columns Before Startup. Chem. Eng.. Apr. 6. 1981. pp. 97-100.
5. Kister. H. Z.. Inspection Assures Troublefree Operation. Chem. Eng.. Feb. 9. 1981. pp. 107-109.
6. Lieberman. N. P.. Basic Field Observations Reveal Tower Flooding. Oil & Gas.J.. Mav 16. 19S8. pp. 39-41.
7. McLaren. D. B.. and Upchurch. J. C.. Guide to Trouble-Free Distillation. Chem. Eng.. June 1. 1970. pp. 139-152.
S. Vital. T.J.. Grossel. S. S.. and Olsen. P. I.. Estimating Separation Efficiency: Part 1: Introduction. Hydrocarbon Proc.. October 1984. pp. 55-56.
Part 2: Packed columns
1. AIChE. "Packed Absorption and Distillation Columns. AIChE Equipment Testing Procedure." American Inst, of Chem. Engrs.. New-York. 1965.
2. Bolles. VV. L.. and Fair. J. R.. Improved Mass-Transfer Model Enhances Packed-Column Design. Chem. Eng.. Julv 12. 1982. pp. 109-116.
2. Chen. G. K.. Packed Column Internals. Chem. Eng.. Mar. 5. 1984. pp. 40-51.
3. Eckert. J. S.. How Tower Packings Behave. Chem. Eng.. Apr. 14. 1975. pp. 70-76.
4. Fadel. T. M.. Selecting Packed-Column Auxiliaries. Chem. Eng.. Jan. 23. 1984. pp. 71-76.
Was this article helpful?